Trapping studies on Glossina longipennis cortie at Nguruman, South Western Kenya

Abstract

Studies have been carried out at Nguruman, south-western Kenya, on Glossina longipennis Corti, a little known member of the fusca group of Glossina. The first objective was to develop an efficient trap suitable for both sampling and control purposes. Studies were then carried out on the population dynamics of G. 1ongipennis using the newly developed sampling methods. Lastly the trap/odour bait system was tested in a control situation. Replicated Latin square design experiments were used to compare the performance of various trap designs and odour attractants. The Zimbabwe F3 trap proved more effective than the widely used biconical trap, especially for females. A new trap developed at Nguruman, called the NG2B, was also very effective and had the advantage of being cheap and easy to construct. Acetone and cow urine together increased the catches by 4-5X over unbaited traps, but when dispensed alone neither of them was effective. There was no significant difference between the attractancy of cow urine and buffalo urine. Trap catches were further increased when l-octen-3-ol was dispensed together with acetone and cow urine. A higher proportion of older flies was caught by the NG2B trap compared to the biconical but no significant difference was observed in the age structure of flies attracted by different odour baits. The effect of trap design on sample composition and the potential for using odour baited traps for sampling the fusca group of tsetse flies are discussed. An electric screen adjacent to a baited target was used to determine the precise activity pattern of G. longipennis which is known to be crepuscular in behaviour. Morning activity started at about 15 minutes before sunrise at 0630 h, peaked at about 0615 h and ceased by 0700 h. The species was more active in the evenings, when activity began at about 30 minutes before sunset at 1815 h, peaked at 1845 h and ended by 1900 h. Males were regularly active before females. Light intensity was found to be the most important factor influencing activity. The relationship between activity pattern and cattle-fly contact is discussed. Changes in the apparent densities of G. longipennis were monitored simultaneously using biconical and NG2B traps in two areas located 7 km apart. Both trap types showed similar trends in population changes but higher apparent densities were recorded with the NG2B trap than with the biconical trap. Apparent densities in both sexes were regularly observed to increase during the rainy seasons and decrease during dry seasons. Peak catches in one area were observed to precede those in the other area by one month. Flies spread out to more open areas during the cool wet seasons and concentrated in the thicker woodland during the dry seasons. The factors influencing changes in population densities, including movement between the two areas and between vegetation t y p e s , are discussed. Mortality rates estimated from ovarian age structure and from Moran curves were observed to be highest during the hot dry seasons and lower during the cool wet seasons. Adult mortality rates showed a significant positive correlation with maximum temperature and a negative correlation with minimum relative humidity. The effect of fly movement on mortality rate estimates and the reliability of the estimates by the two methods are discussed. Dissections of female flies from NG2B traps showed that all non-teneral but nulliparous females and over 80% teneral females were inseminated. The average percentage distribution of the various pregnancy stages in trap samples were found to be very close to the values expected from the duration of the different stages, in contrast to the usual underrepresentation of flies with third instar larvae for other tsetse species. The average abortion rate was 6% but ranged from 0% in the rainy seasons to 60% in the hot dry season. A significant negative correlation was observed between abortion rate and minimum relative humidity. A significant positive correlation was also found between fly size and minimum relative humidity of the previous month but one. A discussion is given of the immediate causes of abortions and their effects on population levels and of the factors influencing fly size. The absolute population size of G. longipennis was estimated through mark-release-recapture experiments. The mean population size was estimated at 17,300 males (range 10,471 - 25,703) and 16,900 females (range 14,125 - 20,892). The trend of changes in the absolute estimates corresponded with those in apparent estimates from trap catches. From the peaks in the recapture rate of marked flies, the feeding cycle of G. longipennis was found to be 2-3 days for males whilst for females the 9-10 day pregnancy cycle was the main factor affecting the recapture rate. There was a considerable amount of movement of marked flies between the two sampling areas but the movement was shown to be greater in one direction than the other. A trial tsetse population suppression operation with baited NG2B traps was started during the course of the study. After 11 months of operation, the population levels of G. 1ongipennis were reduced by about 60% for males and about 90% for females. Much greater reduction levels were obtained for G. pallidipes. A discussion is given of the factors influencing the lesser impact on the population of G. 1ongipennis with suggestions on improving methods for the control of the species.